two notes on the chinese bi comparatives

Concentric: Studies in Linguistics

38.1 (May 2012):69-91

Two Notes on the Chinese bi Comparatives

Chen-Sheng Luther Liu

National Chiao Tung University

Instead of jumping into the debate on whether the Chinese bi comparative is a

clausal or a phrasal comparative, I argue for the following two points concerning this

construction. First, the marker bi is not a degree word; instead, it functions to introduce

the standard and forms a pair of correlative words with a degree adverb with the

meaning of comparison. Second, the standard introduced by bi cannot be elements that

can be modified by a degree adverb; otherwise, a conflict between two dimensions of

comparison will occur.

Key words: bi, correlative word, degree word, dimension conflict

1. Introduction

Kennedy (2005:1) suggests that languages may differ from each other in two

potential parameters in expressing comparison. One is the parameter of explicit versus

implicit comparison and the other is the parameter of individual versus degree

comparison, as shown by (1a-b), respectively.

(1) a. The Parameter of Explicit versus Implicit Comparison

Does comparison involve specialized morphology that expresses arbitrary

ordering relations (explicit comparison), or does comparison involve

taking advantage of the inherent context sensitivity of the positive

(unmarked) form (implicit comparison)?1

Many of the ideas here were intensively discussed with Yi-Hsun Chen, Bo-Ren Hsieh, Hui-Yu

Huang, Hui-Chin Tsai, Ting-Chi Wei and Ching-Yu Yang before they took their present form.

Special thanks to Jo-Wang Lin and Waltraud Paul for their thorough comments and overall critique

of the ideas. I would also like to thank the anonymous reviewers for their insightful and constructive

comments. This research was supported by the research grant (NSC-99-2410-H-009-068) from

National Science Council in Taiwan. Any errors or inconsistencies, of course, are my responsibility. 1 Explicit comparison in English, for example, can be illustrated by (i), which involves the specialized

morphology -er to express the arbitrary ordering relation between the two compared items, John and

Bill along the dimension of height, and implicit comparison by (ii), where the positive form tall is

used.

(i) John is taller than Bill.

(ii) John is tall compared to Bill.

These two different modes of comparison, as Kennedy (2007) argues, differ from each other in that

explicit comparison allows fine-grained distinctions in degree but implicit comparison does not. So,

(i) is grammatically perfect but (ii) is marginally acceptable under the following context.

(iii) Context: John is 5’6” tall and Bill is 5’5 43 ” tall.

38.1 (May 2012)

70

b. The Parameter of Individual versus Degree Comparison

Do comparatives express orderings between arbitrary individuals

(individual comparison), or do they (also) express orderings between

individuals and arbitrary (linguistically explicit) degrees?2

For the individual versus degree comparison distinction, he further suggests the

following two parameters:

(2) The Degree Abstraction Parameter (i.e., DAP) (Beck et al. 2004:325)

A language {does, does not} have binding of degree variables in the syntax.

(3) The Standard Type Parameter

Languages may differ in whether the comparative morphology selects a

standard of type d (degree comparison) or of type e (individual comparison)

Along the line developed by Beck et al. (2004) that only [+DAP] languages allow

the comparative subdeletion, Erlewine (2007), Kennedy (2007), and Xiang (2005)

assume that standards in the Chinese bi comparatives cannot be degree abstraction

structures because the comparative subdeletion is impossible in the Chinese bi

comparative, as shown by the ungrammaticality of (4), taken from Erlewine

(2007:2).3

(4) *Wo-de yizi bi ni-de zhuozi kuan gao.

I-DE chair than you-DE desk wide tall

‘My chair is taller than your desk is wide.’

2 The parameter of individual versus degree comparison can be, respectively, exemplified by (i), which

contains two individual compared items (i.e., John and Bill), and (ii), which has two degrees (i.e., the

degree of width and the degree of tallness) as the compared items.

(i) John is taller than Bill.

(ii) The door is wider than the window is tall. 3 Note that this is not a semantic restriction: (4) as well as example (i) could in principle mean

something like ‘the height of my chair is greater than the width of your desk’.

(i) Wo-de yizi de gaodu bi ni-de zhuozi de kuandu da.

I-DE chair DE height than you-DE desk DE width great

‘The height of my chair is greater than the width of your desk.’

According to Bresnan (1977), in comparative subdeletion structures like the English example (ii), an

amount or degree term must be omitted from the constituent, i.e., the compared constituent (e.g., long

in (ii) that provides the point of comparison with the morphologically marked phrase in the main

clause (e.g., wider in (ii)).

(ii) This desk is wider than that table is (*two inches) long.

In this paper, the term ‘standard’ means the compared constituent introduced by the marker bi, for

example Lisi in (iii), and the term ‘correlate’ means the constituent compared with the standard such

as Zhangsan in (iii).

(iii) Zhangsan bi Lisi gao.

Zhangsan than Lisi tall

‘Zhangsan is taller than Lisi.’

Liu: the Chinese bi comparatives

71

Though agreeing with Beck et al. (2004) and Kennedy (2007) in that the Chinese

bi comparative does not involve degree comparison in syntax, and nor can it be

analyzed as a clausal comparative, Lin (2009) argues that, in addition to the

parameters suggested by Kennedy (2007), two independent parameters are still

needed.4 First, as Lin (2009) points out, the standards (of comparison) in the Chinese

bi comparatives can be normal individuals, times, locations and even propositions, as

long as they are arguments of the predicate of comparison (see the parameter of

argument versus non-argument dependent comparison in (7)).5, 6

(5) a. Zhangsan bi Lisi kaixin.

Zhangsan than Lisi happy

‘Zhangsan is happier than Lisi.’

b. Zhangsan jintian bi zuotian kaixin.

Zhangsan today than yesterday happy

‘Today Zhangsan is happier than yesterday.’

4 A clausal comparative, for example the English than-clausal comparative, is defined as one where the

marker than selects as complement a clause that obligatorily involves the comparative (sub-)deletion.

So, there is no reason to believe that examples like (5d) can only be analyzed as a clausal

comparative. 5 As Lin (2009) argues, there are syntactic reasons to believe that times and locations are more like

arguments than adjuncts with respect to wh-extraction (see Tsai (1994) for discussions on

wh-extraction). Semantically, it is also often assumed, especially in works on tense and aspect, that

time is an argument of a predicate (Lin 2003, 2006). Some linguists even lump time and location

together as one single argument. For example, Kratzer (1988), when discussing the distinction

between stage-level and individual-level predicates, has suggested that stage-level predicates have a

spatial-temporal argument in their argument structure, though these arguments only optionally appear

in overt syntax. 6 Following Liu (1996:220), Lin (2009:25) suggests that the marker bi and the standard(s) introduced

by it form an adjunct constituent adjoined to the left of the predicate of comparison because of the

following evidence. First, a ‘bi standard’ sequence, as (i) illustrates, can be connected with another

one by coordinators like huozhe ‘or’.

(i) Zhangsan bi Lisi huozhe (bi) Wangwu dou haiyao gao.

Zhangsan than Lisi or than Wangwu all even tall

‘Zhangsan is taller than either Lisi or Wangwu.’

Second, a ‘bi standard’ sequence, as (ii) shows, can occur as an independent fragment.

(ii) Ta-de shengao bi wo haiyao ai, bi Yaoming na geng shi

he-DE height than I even short than Yaoming then even.more is

tian cha di yuan le.

heaven differ ground far SFP

‘He is much shorter than I am. If compared with Yaoming, his height is even like the distance

between the heaven and the ground.

Third, as the contrast between (iii) and (iv) indicates, the ‘bi standard’ sequence can only be

‘adjoined’ to the left of the predicate of comparison.

(iii) Zhangsan [ bi Lisi] gao.



(iv) *Zhangsan gao [ bi Lisi].

Zhangsan tall than Lisi

38.1 (May 2012)

72

c. Zhangsan zai jiali bi zai xuexiao kaixin.

Zhangsan at home.inside than at school happy

‘Zhangsan is happier at home than he is in school.’

d. Zhangsan qu bi Lisi lai geng heshi.

Zhangsan to than Lisi come even.more appropriate

‘It is more appropriate for Zhangsan to go there than Lisi to come here.’

This assumption, as Lin (2009) argues, correctly predicts that a manner adverb and a

reason clause cannot occur as the standard because they are not arguments, as the

ungrammaticality of (6a-b) shows.

(6) a. * Zhangsan zhenxinde bi zhengchengde guanxin ni.

Zhangssan sincerely than honestly care.about you

b. * Zhangsan yinwei ni ma ta bi yinwei ni

Zhangsan because you scold he than because you

da ta haiyao shengqi.

beat he even.more angry.

(7) The Parameter of Argument versus Non-argument Dependent Comparison

Comparatives in a language can be argument-dependent or non-argument

dependent comparison.7

So, individual comparison of type e, as Lin (2009) suggests, is just a special case of

argument-dependent comparison.

Second, the number of standards in the Chinese bi comparative, as Lin (2009)

points out, can be one or more than one, as illustrated by examples in (8).

(8) a. Zhangsan bi Lisi kaixin. (one standard)


‘Zhangsan is happier than Lisi.’

7 According to Lin (2009), arguments include individuals, times, locations as well as instruments, but

reasons and manners are adjuncts. So, example (i) with two individual compared items (i.e.,

Zhangsan and Lisi) is a case with argument comparison while (ii) is an example with non-argument

comparison because the compared items are manner adverbs.

(i) Zhangsan bi Lisi gao.



(ii) He signs, but more happily than sadly.


73

b. Zhangsan jintian zai jiali bi Lisi zuotian zai

Zhangsan today at home.inside than Lisi yesterday at

xuexiao kaixin. (three standards)

school happy

‘Zhangsan is happier at home than Lisi was in school yesterday.’

Given this, Lin (2009) suggests that the marker bi is a dyadic comparative degree

word that might take one or more than one argument, either individuals, times,

locations, instruments, or even propositions, as long as they are arguments of the

predicate of comparison. This characteristic of the degree word bi leads Lin (2009) to

propose his second parameter.

(9) The Parameter of Dyadic versus Monoadic Comparison

If a language has phrasal comparatives, the construction may allow only one

phrase (monoadic comparison) or more than one phrase (dyadic comparison)

to be compared.8

Assuming these, Lin (2009) gives example (10a) a syntactic structure like (10b),

in which the degree word bi ‘than’ taking three arguments (i.e., the individual

argument Lisi, the time argument jintian ‘today’, and the location argument zai jiali

‘at home’) must be flanked by constituents of the same type (i.e., the correlate

Zhangsan, zuotian ‘yesterday’ and zai xuexiao ‘at school’). Thus, the whole

DegP-shell headed by the degree word bi ‘than’ occurs as an adjunct degree phrase

adjoined to the predicate of comparison (Tsao 1990).

(10) a. Zhangsan zuotian zai xuexiao [ bi [ Lisi jintian zai

Zhangsan yesterday at school than Lisi today at

jiali ]] kaixin.

home.inside happy

‘Zhangsan was happier yesterday at school than Lisi is at home today.’

8 Example (i) is a case with monoadic comparison because it contains only one pair of compared items,

but (ii) is one with dyadic comparison because three pairs of compared items (i.e., Zhangsan versus

Lisi, jintian ‘today’ versus zuotian ‘yesterday’ and zai jiali ‘at home’ versus zai xuexiao ‘at school’)

are found inside.

(i) Zhangsan jintian bi zuotian kaixin.


‘Today Zhangsan is happier than Lisi was yesterday.’

(ii) Zhangsan jintian zai jiali bi Lisi zuotian zai xuexiao kaixin.

Zhangsan today at home.inside than Lisi yesterday at school happy

‘Today Zhangsan is happier at home than Lisi was at school yesterday.’

38.1 (May 2012)

74

b. [S Zhangsan [AP [NP zuotian] [AP [PP zai xuexiao] [AP [DegP bii [DegP [NP Lisi]

[Deg’ [Deg ti] [DegP [NP jintian] [Deg’ [Deg ti] [PP zai jiali] [AP kaixin]]]]]]]]]].

Semantically, the degree word bi ‘than’ with a denotation like (11), as Lin (2009)

suggests, has a semantic function similar to what Heim (1985) has for the English

comparative morpheme -er in her direct analysis.9

(11) ||bi|| = (l)i(i)

j(w)

kxP<d, <(l), <(i), <(e), <e,t>>>>>(l’)

i(i’)

j(z)

k max d

max d [P(d)(l)(i)(w)(x)]]

Highly influential though it is, Lin (2009) still encounters some problems which

are not easy to cope with. First, assuming that Chinese is a dyadic argument

comparison language and only arguments of the predicate of comparison can be the

standards, Lin (2009) would expect (12a-b), where two reason clauses are compared

with each other, to be ungrammatical, but the fact does not bear out this expectation.10

9 As one anonymous reviewer informed me, it was in Xiang (2005), not Lin (2009), who first proposed

that the word bi heads DegP. Hence, it is better to show the readers how Lin (2009) is similar to and

different from Xiang (2005) with respect to this particular issue. Xiang (2005) proposes a revised

DegP-shell structure for Chinese comparatives by postulating two degree heads in syntactic structure:

one lower than AP and the other above AP, as represented in (i).

(i) [Deg1P [Deg1 bi] [AP standardj [A’ [A k + predicate] [Deg2P standardj [Deg2’ [Deg2 k] [(differential)]]]]]]

Her analysis has two important features. First, it can generate the transitive comparative by

head-movement of the [A k + predicate] to Deg1 when the degree word bi does not appear, as

(ii) illustrates.

(ii) Zhangsan [Deg1P [Deg1 k- gaoi] [AP woj [A’ [A k + gaoi]

Zhangsan tall I tall

[Deg2P woj [Deg2’[Deg2 k][ wu gongfen ]]]]]]

I five centimeter

‘Zhangsan is five centimeters taller than me.’

The other is that the degree morpheme bi and the standard do not form a constituent. Although Lin

(2009) is similar to Xiang (2005) in analyzing the marker bi as a degree word, they still differ from each other in the following ways. First, according to Xiang (2005), the degree word bi can take only

one complement. Second, the complement, as she suggests, is an adjective phrase which further

dominates a degree phrase. Third, Xiang (2005) does not treat the bi-standard sequence as a syntactic

constituent. See Erlewine (2007) and Lin (2009) for further discussion on Xiang (2005). 10 Among the twenty-three native speakers with whom I have checked on (12a-b), all of them accept

(12a) as a natural, understandable and grammatical sentence, and (12b) sounds marginal to nine of

them and grammatical to others. However, as one anonymous reviewer points out to me, “Lin (2009)

assumes traditional adjuncts such as temporal and locative phrases to be arguments; however, the

author didn’t question this assumption. As long as Lin gives an appropriate semantic type to reason

clauses, he can incorporate the apparent ‘counterexamples’ offered by the author into his system

without causing any problem. So, the existence of reason clauses in comparative constructions does

not really stand as a fatal problem to Lin”. Logically, there are three ways to evaluate Lin’s (2009) argument vs. non-argument parameter. One is to point out his (2009) definition of arguments is

incorrect; another is to point out that the argument vs. non-argument parameter based on his

definition of arguments makes wrong predictions empirically; and the other is to point out that Lin’s

(2009) definition of arguments and his argument vs. non-argument parameter are both wrong. Hence,

it is not obligatory for me to question Lin’s (2009) definition of arguments. Thus the comment raised

by the reviewer (i.e., intriguingly, the current author didn’t question this assumption) cannot stand.


75

(12) a. Mama yinwei Xiaomingi shuo huang bi baba

mother because Xiaoming say lie than father

yinwei tai tou dongxi haiyao shengqi.

because he steal thing even.more angry

‘Mother gets angry more because Xiaoming lies than father gets angry

because Xiaoming steals things.’

b. Laoban yinwei ta jingchang chidao bi yinwei ta

boss because he often late than because he

ouer fan cuo hai geng shengqi.

sometimes make mistake even more angry

‘The boss is even more angry because he is often late than because he

sometimes makes mistakes.’

Second, even though all the standards in (13) are arguments of the predicate of

comparison xihuan ‘like’, example (13), which takes the preposed postverbal object

NP of the verb xihuan ‘like’ (i.e., wuli ‘physics’) as one of the standards, is

unexpectedly ill-formed.11

(13) * Zhangsan shuxue bi Lisi wuli xihuan.

Zhangsan mathematics than Lisi physics like

‘Zhangsan likes mathematics more than Lisi likes physics.’

Third, Lin (2009:23) requires the comparative degree word bi not only to be

flanked by standards and their correspondents but also to be of the same type (Tsao

Instead, I simply follow Lin’s (2009) definition of arguments, which regards reasons and manners as

adjunct, and provide grammatical examples where reason clauses occur as compared items to show

that Lin’s (2009) analysis is empirically challenged. Thus, what I have to challenge here is the

already-made proposal in Lin (2009), not what the reviewer says “as long as Lin gives an

appropriate semantic type to reason clauses, he can incorporate the apparent ‘counterexamples’

offered by the author into his system, without causing any problem”. 11

One anonymous reviewer is curious about the origin of the ungrammaticality of (13) and wonders

how its intended meaning can ever be expressed in Chinese. The intended meaning of (13) is

something like ‘the degree that Zhangsan likes mathematics is greater than the degree that Lisi likes

physics’, which can be expressed by examples like (i).

(i) Zhangsan xihuan shuxue de chengdu bi Lisi xihuan wuli de

Zhangsan like mathematics DE degree than Lisi like physics DE

chengdu da.

degree great

‘The degree that Zhangsan likes mathematics is greater than the degree that Lisi likes

physics.’

Since the origin of the ungrammaticality of (13) is beyond the scope of this paper, I will not discuss

it here.

38.1 (May 2012)

76

1990). However, it is not necessary for the degree word bi to be flanked by the

standards and their correlates, as illustrated by (14a-b).12, 13

(14) a. Xianzai wo erzi bi yiqian guaiduo le.

now I son than before well.behaved.more SFP

‘Now my son is more well-behaved than before.’

b. Zai xuexiao xuesheng zongshi bi zai jiali

at school student always than at home.inside

tinghua.

well.behaved

‘Students are more well-behaved in school than at home.’

Moreover, in a Chinese pivotal construction like (15a), the pivotal verb rang ‘let’,

as Tang (2010:184-187) suggests, selects a VP as complement. Hence, it is very

difficult to analyze (15a) as one with a syntactic structure like (15b), where the

predicate zhu san tian ‘live for three days’ takes the string wo rang ni ‘I let you’ as the

sentential subject.

(15) a. Wo [VP [V rang [VP ni [V zhu san tian ]]]].

I let you live three day

‘I let you live here for three days.’

b. [S [S wo rang ni ] [VP zhu san tian ]].

I let you live three day

Given this, Lin’s (2009) analysis will be challenged by (16a), which has a syntactic

structure like (16b) under his analysis, because the degree word bi will take the

non-argumental pivotal verb rang ‘let’ as argument.

12 Among the twenty-three native speakers I have checked on (14a-b), all of them accept (14a-b) as

grammatical sentences. 13 One might say Lin (2009) can avoid this problem by assuming that the temporal adverb xianzai

‘now’ in (14a) is moved from the position adjacent to the bi phrase bi yiqian ‘than before’ to the

sentence-initial position, as (i) shows.

(i) Xianzaii wo erzi ti bi yiqian guaiduo le.

now I son than before well.behaved.much SFP ‘Now my son is more well-behaved than before.’

However, this will bring us the question of why (ii), in which the NP gou ‘dog’ is moved from the

position adjacent to the bi phrase bi mao ‘than cat’, is ungrammatical.

(ii) *Goui Zhangsan ti bi mao xihuan.

dog Zhangsan than cat like

‘Zhangsan likes dogs more than cats.’


77

(16) a. Wo rang ni [ bi ni rang ta ] duo zhu-le san tian.

I let you than you let he many live-ASP three day

‘I let you live here for three more days than you let me.’

b. Wo rang ni [VP [DegP bii [DegP [NP ni ] [Deg’ [Deg ti] [DegP [V rang]

I let you than you let

[Deg’[Deg ti] [NP ta ]]]]]] [VP duo zhu-le san tian ]].

he many live-ASP three day

Although Lin’s (2009) analysis is challenged, I shall not jump into the debate on

whether the Chinese bi comparative is a phrasal or a clausal comparative here (Cheng

1966, Erlewine 2007, Fu 1978, Hashimoto 1966, Kennedy 2007, Liu 1996, Lin 2009,

Paul 1993, Shi 2001, and Xiang 2005). Instead, the purpose of this paper is modest

and straightforward. I want to argue for the following two points in Sections 2 and 3,

respectively. First, the marker bi cannot be analyzed as a degree word; instead, it

simply functions to introduce the standard and forms a pair of correlative words with

a degree adverb with the meaning of comparison. Second, the standard introduced by

the marker bi cannot be elements that can be modified by a degree adverb; otherwise,

a conflict between two dimensions of comparison will occur. It is this conflict that

makes a manner adverb unable to occur as the standard in the bi comparative.

However, as one anonymous reviewer comments, “the overall scope of the paper

is too narrow, and the author should at least comment on the broad theoretical and/or

empirical consequences of the two ‘notes’ made here. Besides, the overall

organization is somewhat unbalanced since the paper puts too much emphasis on the

sort-of side remarks while fails to elaborate on the presumably major proposal.”

Actually, a comprehensive study on the Chinese bi comparative by the author just

came out as Liu (2011), where the space spent on some of the theoretical and/or

empirical consequences of these two notes is not enough. So, this short paper is

written as supplement to that article.

2. The function of bi

According to Lin (2009:19-23), the meaning of comparison of the Chinese bi

comparatives is provided by the word bi, which not only functions to introduce the

standard(s) but also has a function similar to that of the English comparative

morpheme -er. So, the marker bi is analyzed as a comparative degree word and the

phrase headed by it is a degree phrase. This analysis, however, will encounter the

following empirical and theoretical problems. First, in Chinese, an adjunct degree

38.1 (May 2012)

78

modifier must precede and be adjacent to the gradable constituent modified by it, as

illustrated by the contrast below.14

(17) a. Wo dui ni hen keqi.

I to you very courteous

‘I am very courteous to you.’

b. * Wo hen dui ni keqi.

I very to you courteous

Suppose the marker bi is a comparative degree word, the degree phrase headed by it

and the gradable constituent modified by it should not be intervened by any syntactic

constituent. However, as the grammaticality of (18) shows, the fact does not bear out

this prediction.

(18) Zhangsan bi Lisi dui wo keqi.

Zhangsan than Lisi to I courteous

‘Zhangsan is more courteous to me than Lisi is.’

Second, a degree phrase cannot occur in a position hierarchically higher than a

locative phrase, as the contrast below illustrates.

(19) a. Zhangsan zai meiguo geng chidekai.

Zhangsan at America even.more influential

‘Zhangsan is even more influential in America.’

b. * Zhangsan geng zai meiguo chidekai.

Zhangsan even.more at America influential

Thus, it is expected that a bi phrase, if being analyzed as a degree phrase, cannot

occur in a position hierarchically higher than a locative phrase. However, this

expectation is not borne out, as (20) shows.

14

One might challenge this analysis by pointing out that a negation marker like bu ‘not’ may be

inserted between a degree modifier and the gradable predicate modified, as shown below.

(i) Ni zheyang zuo tai bu yinggai.

you this.way do too not should

‘It is too inappropriate for you to do in this way.’

However, this claim can hold only if the negation marker bu ‘not’ in (i) cannot be analyzed as a

lexical negation marker. Indeed, it is very difficult to exclude this possibility.


79

(20) Zhangsan bi Lisi zai meiguo chidekai.

Zhangsan than Lisi at America influential

‘In America, Zhangsan is more influential than Lisi.’

These points lead us to propose the following alternative analysis: The word bi

simply functions to introduce the standard(s) and the meaning of comparison in the bi

comparative is provided by the comparative degree adverb like geng ‘even more’ or

the covert comparative morpheme (i.e., ), as shown by (21a-b), respectively.15, 16

(21) a. Zhangsan bi Lisi geng gao.

Zhangsan than Lisi even.more tall

‘Zhangsan is even taller than Lisi.’

b. Zhangsan bi Lisi gao.

Zhangsan than Lisi GENG tall


The supporting evidence for the existence of the covert comparative morpheme comes

from the fact that a Chinese gradable adjective can occur as predicate only if it

co-occurs with a degree term (Liu 2010). So, the impossibility of treating the marker

bi as a degree word leads us to suggest that, in the Chinese bi comparatives with an

adjectival predicate, a comparative morpheme, either the overt geng ‘even more’ or

the covert , is obligatorily required. Furthermore, a Chinese bi comparative with

15 To get the meaning of comparison for examples like (21b), one might resort to the suggestion made

by Sybesma (1999:26-27) and Xiandai Hanyu Xuci Lishi (1982:243-244); that is, Chinese adjectives

differ from their European counterparts in that the latter choose the unmarked option for the positive

degree but the former the unmarked option for the comparative. Namely, in European languages the

comparative is morphologically marked whereas in Chinese the positive degree is marked by the

most neutral ‘positive degree marker hen’, as shown by the contrast below.

(i) a. John is taller. b. John is tall.

(ii) a. ( Zhangsan han Lisi, shei gao?) Zhangsan gao.

Zhangsan and Lisi who tall Zhangsan tall

‘(As for Zhangsan and Lisi, who is taller?) Zhangsan is taller.

b. Zhangsan *( hen ) gao.

Zhangsan HEN tall

‘Zhangsan is tall.’

However, this suggestion is not without problems because the language fact is more complex than

what Sybesma (1999) and Xiandai Hanyu Xuci Lishi (1982) indicate. As Zhu (1982) and Liu et al.

(2004) point out, in Chinese it is possible for a positive-degree-denoting ‘unmarked’ gradable

adjective to occur as predicate, for example, in the bu ‘not’ negation sentence, the contrastive focus

construction, the ma particle question, the epistemic adjectival small clause, and the conditional. Please see Liu (2010) for further discussion.

16 One anonymous reviewer reminds me that it might be methodologically problematic to say that it is

the only possibility left that the marker bi simply functions to introduce the standard and forms a

pair of correlative words with a degree adverb with the meaning of comparison. Actually, I have no

intension of making such a strong claim; therefore, I use ‘the following alternative analysis’ to

replace my original wording ‘the suggestion’ to avoid confusion.

38.1 (May 2012)

80

an adjectival predicate must contain a bi-phrase if no overt comparative morpheme

(i.e., the overt degree adverb geng ‘even more’) is found; in other words, if what

occurs in a Chinese bi comparative with an adjectival predicate is the covert degree

adverb geng, then the bi-phrase is obligatorily required. Thus, the occurrence of the bi

phrase in a bi comparative with an adjectival predicate implies the occurrence of a

comparative morpheme and vice versa. Thus, I suggest that the word bi and the

comparative degree morpheme form a pair of correlative words in a Chinese bi

comparative with an adjectival predicate.

The correlative relation between the word bi and the comparative degree adverb in

the bi comparative naturally explains why we cannot understand (22) as a

comparative construction with the covert comparative morpheme.

(22) Ni qingchu-le ma?

you clear-ASP SFP

‘Is it clear to you?’

Since the occurrence of the covert comparative morpheme is possible only if

the sentence contains the word bi, example (22), which does not contain the marker bi,

cannot get the meaning of comparison.

3. A restriction on the standard(s)

As pointed out above, the standard in the Chinese bi comparative can be

individuals (NPs), times (NPs or PPs), locations (PPs), instruments (VPs or PPs),

propositions (CPs), or even reasons (CPs or PPs) (Beck et al. 2004, Heim 1985,

Kratzer 1988, Kennedy 2007, and Lin 2009).

(23) a. Zhangsan bi Lisi kaixin. (individual)


Zhangsan is happier than Lisi.’

b. Zhangsan jintian bi zuotian kaixin. (time)


‘Today Zhangsan is happier than Lisi was yesterday.’

c. Zhangsan zai xuexiao bi zai jiali tinghua. (location)

Zhangsan at school than at home.inside obedient

‘Zhangsan is more obedient in school than Lisi is at home.’


81

d. Zhangsan yong chazi bi yong kuaizi shoulian. (instrument)

Zhangsan use fork than use chopstick skilled

‘Zhangsan is more skilled in using forks than using chopsticks.’

e. Laoban yinwei ta jingchang chidao bi yinwei ta

boss because he often late than because he

ouer fan cuo hai geng shengqi. (reason)

sometimes make mistake even more angry

‘The boss is even more angry because he is often late than because he

sometimes makes mistakes.’

f. Ni qu bi wo qu heshi. (proposition)

you go than I go appropriate

‘It is more appropriate for you to go than for me to go.’

Importantly here, the fact that adjuncts like reason clauses, as (23e) shows, can occur

as the standard in the Chinese bi comparative makes Lin’s (2009) parameter of

argument versus non-argument dependent comparison questionable. To answer the

question of why a manner adverb cannot occur as the standard, I shall argue that the

occurrence of a ‘manner adverb’ standard in the bi comparative will arouse a conflict

between two dimensions of comparison, which makes a bi comparative

ungrammatical.

Here, I simply follow Lin’s (2009) analysis on the bi comparative that involves

one standard. The semantics of the bi phrasal comparative, as Lin (2009) suggests,

can be generated by Heim’s (1985) direct analysis of phrasal comparatives. As Heim

(1985) suggests, the general meaning of -er can be specified as (24) and what are

compared in the phrasal comparative, for example, can be two individuals.

(24) “-er <a, b>f” is true iff f(a) > f(b).

So, a function from individuals to degrees which can be represented by a lambda-iota

expression “xy[…x…y…]” (i.e., f in (24)) is needed. This implies that what are

compared in the phrasal comparatives are preferred to be elements without carrying

degrees along some dimension by themselves; otherwise, the function of f in (24)

would be nullified.

Crucially here, in the bi comparatives, the dimension of comparison is provided

by the predicate modified by the degree adverbs with the comparison meaning such as

geng ‘even more’ (or, more precisely, the predicate that occurs after the adverb with

38.1 (May 2012)

82

the comparison meaning in the linear order).17

Furthermore, this predicate, as

illustrated in (25a-b), cannot contain the standard.

(25) a. Zhangsan [[ bi Lisi] [ geng [ gao]]].

Zhangsan than Lisi even.more tall

‘Zhangsan is even taller than Lisi.’

b. Zhangsan [[ feidan [ bi Lisi]] [ erqie [ bi Wangwu]]]

Zhangsan not.only than Lisi but.also than Wangwu

[[ geng [[ chang] lai kan ni ]]].

even.more often come visit you

‘Zhangsan visits you more often than Lisi and Wangwu.’

Given this, if the standard is gradable by itself, the dimension provided by it and that

provided by the predicate of comparison will conflict with each other. Therefore, we

would expect example (26a-c) to be ungrammatical, and the fact bears out this

expectation.

(26) a. * Zhangsan [ jingchangde [[ bi ouerde ] [ geng guanxin

Zhangsan often than sometimes even.more care.about

ta]]].

he

b. * Zhangsan[ gongkaide [[ bi sidixiade]] [ geng hui shuo ren

Zhangsan in.public than in.private more able say people

huai hua]]].

bad word

c. * Zhangsan [ zhenxinde [[ bi zhenchengde] [ geng guanxin

Zhangsan sincerely than genuinely even.more care.for

ni ]]].

you

As (27a-c) show, the temporal adverb jingchangde ‘often’ and manner adverbs like

gongkaide ‘publicly’ and zhenxinde ‘sincerely’ are all gradable because they can be

modified by a degree adverb.

17

The dimension of comparison might be provided by the head of the predicate, for example gao ‘tall’

in (25a) or a gradable adverbial modifier contained in the predicate like chang ‘often’ in (25b).


83

(27) a. Zhangsan hui geng jingchangde lai tanwang ni.

Zhangsan will even.more often come visit you

‘Zhangsan will even more often visit you.’

b. Zhangsan geng gongkaide piping ni.

Zhangsan even.more in.public criticize you

‘Zhangsan even more publicly criticizes you.’

c. Zhangsan geng zhenxinde guanhuai ni.

Zhangsan even.more sincerely care.about you

‘Zhangsan even more sincerely cares about you.’

Thus, the question of why manner adverbs and temporal adverbs like jingchangde

‘often’ cannot occur as standards is naturally explained.18

In other words, the crucial factor that makes a manner adverb and adverbs like

jingchangde ‘often’ unable to occur as the standard in the Chinese bi comparative is

that the standard introduced by the marker bi cannot be gradable (i.e., elements that

can be modified by a degree adverb) and the dimension of comparison can only be

provided by the predicate occurring behind the degree adverb with comparison

meaning like geng ‘even more’ in the linear order (e.g., guanxin ta ‘care about him’ in

(26a)). However, this restriction, as illustrated by (28a-c) taken from Lin (2009:17),

does not exist in the English comparatives.

(28) a. Although there may be a $100 million picture out there, I think that its

day will [[come more slowly] [than quickly]], so that the market will be

allowed a chance to [[grow more slowly] [than swiftly]].

b. In this sense, cultural organizations – for the most part public or at lease

dependent on public funding – take part, more reluctantly than willingly,

in this process of ….

c. He signs, but more happily than sadly.

That is, the English comparatives allow the correlate to be gradable and to take the

comparative morpheme more/-er. Since the dimension of comparison is provided by

18

One anonymous reviewer points out to me that (26b) is perfectly fine as long as de is removed from

the adverbs, as shown below.

(i) Zhangsan[ gongkai [[ bi sidixia ]] geng hui shuo ren huai hua ]]].

Zhangsan in.public than in.private even.more able say people bad word

‘Zhangsan maliciously makes more remarks on people on a public occasion than on a private

occasion.’

I have checked nine native speakers on (i), and they all say (i) is acceptable only if gongkai ‘public’

and sidixia ‘private’ are, respectively, understood as zai gongkai de changhe ‘on a public occasion’

and zai sidixia de changhe ‘on a private occasion’. In other words, (i) is acceptable only if gongkai

‘public’ and sidixia ‘private’ are understood as locations rather than manner adverbs.

38.1 (May 2012)

84

the syntactic element modified by the comparative morpheme (e.g., slowly in (28a)) in

the English comparatives, no conflict between dimensions of comparison occurs in

(28a), where the standard swiftly has slowly as correlate. This analysis is further

supported by correctly ruling out examples like (29).

(29) *John is more concerned about you sincerely than kindly.

In (29), both the standard (i.e., kindly) and its correlate (i.e., sincerely) are gradable,

but the comparative morpheme more is attached to the adjectival predicate concerned.

As a result, a conflict between two dimensions of comparison occurs. So, (29) is

ungrammatical.

However, one anonymous reviewer questions my assumption that the standard

introduced by bi cannot be an element that can be modified by a degree adverb by

pointing out that the standard introduced by the marker bi can be gradable, as (30)

illustrates (i.e., yonggan ‘brave’).

(30) Qinlao bi yonggan geng zhongyao.

diligent than brave even.more important

‘Diligence is even more important than bravery.’

This example, though at first glance it seems like a counterexample, actually provides

supporting evidence for my assumption. Since the morphological difference between

brave and bravery in English does not overtly show up in Chinese, it is not

implausible for us to say that in Chinese the adjective yonggan with the meaning of

brave and the noun yonggan with the meaning of bravery are homophones. So, the

word yonggan in (30) can be analyzed as the noun yonggan ‘bravery’.19

Since a noun

like yonggan ‘bravery’ cannot be modified by a degree adverb, (30), as predicted, is

grammatical. Supporting evidence for this account comes from the following two

facts. First, as (31a-b) show, yonggan ‘bravery’, when occurring as subject, can be

replaced by the pronominal zhe ‘this’.

(31) a. Yonggan shi yizhong renge tezhi.

bravery is one.kind personal characteristic

‘Bravery is a kind of personal characteristics.’

19

More precisely, here I treat yonggan ‘bravery’ as a noun rather than a process nominal to deal with

the question of why it cannot occur with an adverb (Fu et al. 2001).


85

b. Zhe shi yizhong renge tezhi.

this is one.kind personal characteristic

‘This is a kind of personal characteristics.’

Second, example (30) can be rewritten as (32) without any significant change in

meaning.

(32) Zhe bi na geng zhongyao.

this than that even.more important

‘This is even more important than that.’

However, as the reviewer also points out, treating yonggan in (30) as a noun

might be challenged by examples like (33a-c), in which shaotangshang ‘scald’,

shaoshang ‘burn’, jusang ‘depressed’, beishang ‘sad’, jueshi ‘hunger strike’ and jieshi

‘diet’ are all nominalized elements and occur with a degree adverb. (Among the seven

native speakers I consulted, five of them consider (33c) as being marginal and the

other two consider it as being ungrammatical.)

(33) a. Qingdu shaotangshang bi yanzhong shaoshang

mild scald than serious burn

xingyun duo le.

lucky more SFP

‘It is more lucky for one to be mildly scalded than to be

seriously burned.’

b. Jidu jusang bi jidu beishang hao duo le.

extremely depressed than extremely sad good more SFP

‘It is better for one to be extremely depressed than to be extremely sad.’

c. Jueshi bi shaowei jieshi kepa.

hunger.strike than slight diet terrible

‘A hunger strike is more terrible than a mild diet.’

The reviewer’s further comments can be briefly summarized as the following three

points: (i) degree words include feichang ‘extremely, yanzhong ‘seriously’, qingdu

‘mildly’, jidu ‘extremely’, shaowei ‘slightly’ and tebie ‘especially’; (ii) a nominalized

element can be modified by a degree adverb; and (iii) Chinese does not allow multiple

degree adverbs in a row.

Unfortunately, the reviewer’s comments are self-contradictory. Following the

reviewer’s comment (ii) (i.e., “a nominalized element can be modified by a degree

38.1 (May 2012)

86

adverb), if words like qingdu ‘mild’, yanzhong ‘serious’, jidu ‘extreme’ and shaowei

‘slightly’ in (33a-c), as the reviewer suggests in comment (i), are all degree adverbs,

then example (34a-c) will be incorrectly ruled out by the reviewer’s comment (iii) (i.e.,

Chinese does not allow multiple degree adverbs in a row) because, as (35) shows, it is

extremely difficult for the reviewer to say that words like feichang ‘extremely’, jiwei

‘extremely’ and youdian ‘slightly’ are not degree adverbs.20

(34) a. Buguan zenmeyang, feichang yanzhongde shaoshang haishi

no.matter how extremely serious burn still

bi feichang yanzhongde cashang nan zhiliao.

than extremely serious scrape difficult treat

‘No matter what it is, an extremely serious burn is more

difficult to treat than an extremely serious scrape is.’

b. Buguan zenmeyang, jiwei jidude tangshang haishi

no.matter how extremely extreme scald still

bi jiwei qingdude cashang nan zhiliao.

than extremely mild scrape difficult treat

‘No matter what it is, an extremely serious scald is more

difficult to treat than an extremely mild scrape is.’

c. Buguan zenmeyang, shaowei youdian beishang haishi bi

no.matter how slight slight sad still than

shaowei youdian youyu hao.

slight slight depressed good

‘No matter what it is, a little bit slight sadness is better than a

little bit slight depression is.’

(35) Zhangsan feichang/jiwei/youdian beishang.

Zhangsan extremely/extremely/slightly sad

‘Zhangsan is extremely/extremely/slightly sad.’

Although I agree with the reviewer that “a nominalized element can occur with an

adverb”, this does not mean we can say that all types of nouns (or nominal elements)

can occur with an adverb or all types of adverbs can occur with a noun (or a nominal

element) (Fu et al. 2001). Thus, the fact that “a nominalized element can occur with

20 Perhaps, the reviewer might say in (34a-c) the first degree adverb occurs as a modifier for the

second degree adverb; for example, in (34a) what is modified by the degree adverb feichang

‘extremely’ is the degree adverb yanzhongde ‘seriously’ rather than the expression yanzhongde

shaoshang ‘serious burn’. This assumption, however, is immediately challenged by the

ungrammaticality of the reviewer’s examples like (i).

(i) *Zhangsan feichang tebie gao.

Zhangsan extremely especially tall


87

an adverb” does not guarantee that the nominalized element that follows the marker bi

can be modified by a degree adverb. Consider the following examples.

(36) a. qingdu(-de) taifeng

mild(-DE) typhoon

‘a mild typhoon’

b. yanzhong(-de) bingzheng

serious(-DE) disease

‘a serious disease’

c. jidu(-de) qingyu

extreme(-DE) lust

‘the extreme lust’

There is no way to exclude the possibility of analyzing qingdu ‘slight’, yanzhong

‘serious’ and jidu ‘extreme’ in (36a-c) as an attributive adjectival modifier. If the

word yanzhong ‘serious’, qingdu ‘mild’ and jidu ‘extreme’ in (33a-b) and (34a-b) are

treated as an attributive adjectival modifier, then the grammaticality of (34a-b) are

expected.

Also, it is not so certain to say that the word shaowei ‘slightly’ in (33c) can only

be analyzed as a degree adverb. When being used as a degree adverb, shaowei

‘slightly’, which must syntactically co-occur with the measure phrase yidian ‘a little

bit’, can only express the explicit comparison, while youdian ‘a little bit’ the implicit

comparison, as the contrast between (37a-b) and (38a-b) illustrates (Kennedy 2007).21

(37) a. Zhangsan bi Lisi shaowei gao *( yidian ).

Zhangsan than Lisi slightly tall a.little.bit

‘Zhangsan is a little bit taller than Lisi.’

b. Zhangsan shaowei gao *( yidian ).

Zhangsan slightly tall a.little.bit

‘Zhangsan is a little bit taller.’

(38) a. *Zhangsan bi Lisi youdian gao ( yidian ).

Zhangsan than Lisi a.little.bit tall a.little.bit

‘Zhangsan is a little bit taller than Lisi.’

b. Zhangsan youdian gao *( yidian ).

Zhangsan a.little.bit tall a.little.bit

‘Zhangsan is a little bit tall.’

21

The co-occurrence restriction between the degree adverb shaowei ‘slightly’ and the measure phrase

yidian ‘a little bit’ explains why (33c) sounds marginal to five of my seven informants and

ungrammatical to the other two.

38.1 (May 2012)

88

Here relevant to the contrast above is the grammaticality of example (39a), in which

the adverb shaowei ‘slightly’ co-occurs with youdian ‘a little bit’.

(39) a. Zhangsan shaowei youdian qiong.

Zhangsan slightly a.little.bit poor

‘Zhangsan is a little bit poor.’

b. * Zhangsan bi Lisi shaowei youdian qiong.

Zhangsan than Lisi slightly a.little.bit poor

Since an implicit-comparison-denoting adverb cannot co-occur with an explicit-

comparison-denoting adverb, we would expect (39a) to be ungrammatical if both

shaowei ‘slightly’ and youdian ‘a little bit’ can only be used as degree adverbs;

however, the fact does not bear out this expectation. Moreover, as the meaning of (39a)

shows, the word shaowei ‘slightly’ in (34c) should not be analyzed as a degree

adverb.

4. Concluding remarks

Since the syntax and semantics of the Chinese bi comparative is extremely

complicated, my aim in this paper is modest. I tentatively conclude that the marker bi

is not a degree word with the meaning of comparison; instead, it functions to

introduce the standard(s) and forms a pair of correlative words with the adverb with

the comparison meaning like geng ‘even more’ in the Chinese bi comparative. The

standard introduced by the marker bi cannot be elements that can be modified by a

degree adverb. Otherwise, a conflict between two dimensions of comparison would

occur.

References

Beck, Sigrid, Toshiko Oda, and Koji Sugisaki. 2004. Parametric variation in the

semantics of comparison: Japanese vs. English. Journal of East Asian Linguistics

13.4:289-344.

Bresnan, Joan. 1977. Variables in the theory of transformations. Formal Syntax, eds.

by Peter Culicover, Thomas Wasow, and Adrian Akmajian, 157-196. New York:

Academic Press.

Cheng, Robert L. 1966. Some Aspects of Mandarin Syntax. Doctoral dissertation,

Indiana University, Bloomington.


89

Erlewine, Michael. 2007. A New Syntax-Semantics for the Mandarin bi Comparative.

MA thesis, University of Chicago, Chicago.

Fu, Jingqi, Thomas Roeper, and Hagit Borer 2001. The VP within process nominals:

evidence from adverbs and the VP anaphor do-so. Natural Language and

Linguistic Theory 19.3:549-582.

Fu, Yi-Chin. 1978. Comparative Structures in English and Mandarin Chinese.

Doctoral dissertation, University of Michigan, Ann Arbor.

Hashimoto, Anne. 1966. Embedding Structures in Mandarin. Doctoral dissertation,

Ohio State University, Columbus.

Heim, Irene. 1985. Notes on Comparatives and Related Matters. Austin: University of

Texas, Austin, manuscript.

Kennedy, Christopher. 2005. Parameters of Comparison. Chicago: University of

Chicago, manuscript.

Kennedy, Christopher. 2007. Modes of comparison. Chicago Society of Linguistics

43.1:141-165.

Kratzer, Angelika. 1988. Stage-level and individual-level predicates. Genericity in

Natural Language, ed. by Manfred Krifka, 247-284. Tübingen: University of

Tübingen.

Lin, Jo-Wang. 2003. Temporal reference in Mandarin Chinese. Journal of East Asian

Linguistics 12.4:259-311.

Lin, Jo-Wang. 2006. Time in a language without tense: the case of Chinese. Journal

of Semantics 23.1:1-53.

Lin, Jo-Wang. 2009. Chinese comparatives and their implicational parameters.

Natural Language Semantics 17.1:1-27.

Liu, Chen-Sheng Luther. 1996. A note on Chinese comparatives. Studies in the

Linguistics Science 26.1/2:215-235.

Liu, Chen-Sheng Luther. 2010. The positive morpheme in Chinese and the adjectival

structure. Lingua 120.4:1010-1056.

Liu, Chen-Sheng Luther. 2011. The Chinese bi comparative. Lingua

121.12:1767-1795.

Liu, Yue-Hua, Wen-Yu Pan and Hua Gu. 2004. Shiyong Xiandai Hanyu Yufa

[Modern Chinese Grammar]. Beijing: Shangwu Yinshuguan.

Paul, Waltraud. 1993. A non-deletion account of the comparative construction in

Mandarin Chinese. Cahiers de Linguistique Asie Orientale 22.1:9-29.

Shi, Ding-Xu. 2001. The nature of Chinese comparatives. Studies in Chinese

Linguistics vol. 2, ed. by Hai-Hua Pan, 137-158. Hong Kong: Linguistic Society

of Hong Kong.

Sybesma, Rint. 1999. The Mandarin VP. Dordrecht: Kluwer Academic Publishers.

38.1 (May 2012)

90

Tang, Sze-Wing. 2010. Xingshi Hanyu Yufa [Formal Chinese Syntax]. Shanghai:

Shanghai Educational Publishing House.

Tsai, Wei-Tien Dylan. 1994. On nominal islands and LF extraction in Chinese.

Natural Language and Linguistic Theory 12.2:121–175.

Tsao, Feng-Fu. 1990. Sentence and Clause Structure in Chinese: A Functional

Perspective. Taipei: Student Book Co.

Xiandai Hanyu Xuci Lishi [Examples and Explanation of the Functional Words of

Modern Chinese]. 1982, eds. by Class 1955 and 1957 of Department of Chinese

Language and Literature of Peking University. Beijing: Shangwu Yinshuguan.

Xiang, Ming. 2005. Some Topics in Comparative Constructions. Doctoral dissertation,

Michigan State University, East Lansing.

Zhu, De-Xi. 1982. Yufa Jiangyi [Lectures on Chinese Syntax]. Beijing: Shangwu

Yinshuguan.

[Received April 11, 2011; revised November 11, 2011; accepted December 16, 2011]

Department of Foreign Languages and Literatures

National Chiao Tung University

Taipei, TAIWAN

[email protected]

mailto:[email protected]


91

有關漢語「比字比較句」的兩個註記

劉辰生

國立交通大學

本文在不涉入「比字比較句」應否分析為「子句比較句」或是「詞

組比較句」的前提下，論證此一結構具有下列兩個語法及語意特性：

第一、「比」字不是一個程度詞；它的句法作用在引介「比較標準項」

並與之組成一個加接於比較謂語左側的附加語詞組；在語意上「比」

字會和句中表「比較」語意的程度副詞組成一對關聯詞。第二、「比」

字所引介的「比較標準項」不能是一個可以受程度副詞修飾的詞組，

否則會引發「向度衝突」，從而造成句子的不合語法。

關鍵詞：「比」、關聯詞、程度詞、向度衝突

two notes on the chinese bi comparatives

Documents